search

historical-data / schema

Microdata schema for historical data.
historical-data.org
30 stars 4 forks source link

remove the deceased flag #8

Closed stoicflame closed 12 years ago

stoicflame commented 12 years ago

Remove the deceased flag (which can be inferred from the existence of a death event or death date) to reduce redundancy.

ninjudd commented 12 years ago

A separate deceased property is included because there may not be any information about the death event. In this case, adding an empty death event is cumbersome.

stoicflame commented 12 years ago

A separate deceased property is included because there may not be any information about the death event. In this case, adding an empty death event is cumbersome.

You might be right.

It's just that now you have to deal with the inconsistency questions such as "what if there's a death date, but the deceased property is "false"?

And why would search engines and web crawlers even need to know whether a person is dead? Users are usually searching for death date anyway.

ninjudd commented 12 years ago

Undead?

stoicflame commented 12 years ago

Undead?

LOL!

ninjudd commented 12 years ago

Again this raises the question of how accurately we want the schema to represent the underlying data. If you're right that deceased is not important for search engines, then I'd be okay removing it.

RobertGardner commented 12 years ago

The real question isn't "do search engines need it?" but "what types of searches will people be doing?"

I don't do enough genealogical searches to answer that fully, but it seems to me that people wouldn't search for "find Robert Gardner who is dead" but would instead search for "find Robert Gardner who died around 1850".

I would recommend, though, that if we can't find a way to indicate "dead but not sure when" in the death event, then we should insert an empty record. Or find a way to modify the event to indicate the date is unknown.

NatAtGeni commented 12 years ago

I suppose it would be workable to create dates with simply 'unknown' for the date - this wouldn't apply only to death events, but to any event.

ninjudd commented 12 years ago

Couldn't we just create an empty Event with no date specified?

stoicflame commented 12 years ago

I would recommend, though, that if we can't find a way to indicate "dead but not sure when" in the death event, then we should insert an empty record. Or find a way to modify the event to indicate the date is unknown.

Why? Can someone articulate the use case where this information would be useful?

RobertGardner commented 12 years ago

What do genealogy sites do now when the death date is unknown? Don't they put something like "Death: Unknown"? If so, then why not include that in the microdata? If it is a difficult process to provide the information, then I agree it may not be worth the effort, but some of the proposals here sound simple and natural.

What do we do with birth dates that are unknown? Shouldn't we treat them both the same?

ninjudd commented 12 years ago

That was the idea behind the original deceased field. With that in place, you can just treat birth and death the same and leave them blank for unknown data.

stoicflame commented 12 years ago

I'm having a hard time assessing where we landed on this. I'm okay keeping the deceased flag if we think it will help in search results. Should I close the issue or do we need more discussion?

RobertGardner commented 12 years ago

I commented on this thread to express my personal preferences, but I don't have a strong stance on this. I could go either way. I just think that it would be a good idea to have a means of explicitly stating the person is deceased but we don't know the date. Personally, I thought the UNKNOWN date or some variant of it (an event with no date?) was better than the deceased field.

I also worry that if we have a deceased field and people regularly put it there with value "false" then a generic search engine that doesn't understand the schema might index "deceased" on the person, resulting in confusing search results. So if we keep the "deceased" field we should document that it should only be present if the person is actually deceased.

Just realized that we might want to suggest people not include any attributes for events that aren't known to have happened (if that's not obvious already). For example, don't include a christening event with UNKNOWN unless we know the person was christened but we just don't know the date.

ninjudd commented 12 years ago

Let's go with an empty death to indicated deceased with an unknown date and remove deceased.